Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Defining a radiating surface without a solid

Please login with a confirmed email address before reporting spam

Hi all,

I need to define a radiating surface at fixed temperature that "illuminates" my model. I am not interested in how the heat gets to the surface, so I would like to avoid modeling a solid to "support" the surface, but I don't know hot to do it. In fact, if I model the surface as just a surface, I cannot include it in the Heat Transfer in Solids module, since it is not part of a domain.

Reasons to want to avoid modeling a solid:
- reduce complexity of the model
- not worry about radiating/shading effect of the other surfaces of the solid
While there are certainly workarounds to this issues (that can, however, get tricky in complicated geometries), it is the general principle that interests me. Is it possible to define a surface as a radiation source, without having a corresponding domain included in the heat transfer simulation?


Is this possible?

Thanks

Giacomo

6 Replies Last Post 28.08.2015, 15:10 GMT-4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 26.08.2015, 03:56 GMT-4
Hi

but, in all generality it's not that simple: with radiation you have the mutual view factors, emissivity and it's spatial distribution and the spectral response that might interfere and change substantially your incident radiation power.

The simplest in COMSOL is to say you exchange with an infinite sized radiation source (view factor =1) of given temperature you need the HT module and to turn on radiative exchange i.e. Radiation - diffuse surface.

another way is to add just the estimated heat flux onto your boundaries, but then you must write out the equation if you want to be T^3 temperature dependent


Check the HT doc under radiation

--
Good luck
Ivar
Hi but, in all generality it's not that simple: with radiation you have the mutual view factors, emissivity and it's spatial distribution and the spectral response that might interfere and change substantially your incident radiation power. The simplest in COMSOL is to say you exchange with an infinite sized radiation source (view factor =1) of given temperature you need the HT module and to turn on radiative exchange i.e. Radiation - diffuse surface. another way is to add just the estimated heat flux onto your boundaries, but then you must write out the equation if you want to be T^3 temperature dependent Check the HT doc under radiation -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 26.08.2015, 10:50 GMT-4
Hi Ivar,

thanks for the answer... but I'm not sure I understand it.

First of all, your answer starts with a "but", that makes me think there was another statement before the "but" that got lost somehow?

In any case, I'm not sure what your point is: all the problems you mention are either taken care of already in COMSOL (view factor, emissivity) or an issue of how you want to model your physical system. In my case I don't care about spectral response, for example.

What I'm asking here is not a physical question, but a COMSOL one: suppose I have a geometry, and I want to calculate the heat distribution when I "illuminate" it with a uniform radiating surface of a specific size in a specific location. I can do this by adding to the model a box one of the face of which corresponds to my radiating surfaces. The problem is that now I have to deal with another 5 faces and a domain, that I don't need to (and actually would rather not have to) include in the simulation.

Rather than a COMSOL model, let me attach a picture illustrating the problem using an hypothetical simplified 2D model. Left is what I would like to obtain. Right is what I need to now, since I don't know how to include a radiating surface without it being part of a closed domain.

Hope this is clearer now.

Giacomo
Hi Ivar, thanks for the answer... but I'm not sure I understand it. First of all, your answer starts with a "but", that makes me think there was another statement before the "but" that got lost somehow? In any case, I'm not sure what your point is: all the problems you mention are either taken care of already in COMSOL (view factor, emissivity) or an issue of how you want to model your physical system. In my case I don't care about spectral response, for example. What I'm asking here is not a physical question, but a COMSOL one: suppose I have a geometry, and I want to calculate the heat distribution when I "illuminate" it with a uniform radiating surface of a specific size in a specific location. I can do this by adding to the model a box one of the face of which corresponds to my radiating surfaces. The problem is that now I have to deal with another 5 faces and a domain, that I don't need to (and actually would rather not have to) include in the simulation. Rather than a COMSOL model, let me attach a picture illustrating the problem using an hypothetical simplified 2D model. Left is what I would like to obtain. Right is what I need to now, since I don't know how to include a radiating surface without it being part of a closed domain. Hope this is clearer now. Giacomo


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 26.08.2015, 13:29 GMT-4
Hi

Probably I had a first sentence in my head, or it was a but ... from the middle of your long question ;)

I see two/three cases:
1) your "illuminator" is infinite and opaque or covers 2*pi str above you illuminated area and your emissivity and diffusivities are fixed, known, or just =1 => you know the incident flux W/m^2 and you give an incident power density on the illuminated surface, and you forget about the illuminator domain

2) Or your illumination surface is limited to an area A_illu and you simplify the case and say the power emitted is fully absorbed, uniformly, (or partially if you can justify how much) over the illuminated Area and you say you have an incident flux density of the emitted power corrected for the relative Area ratio, and you forget about the illuminator domain too.

3) Or you have to take into account View Factors, hence solid angles, distances between surfaces, angles, Areas, emissivity, scatter ... And for that you must define the domain between the illuminator to the illuminated.

still, if your illuminated domain is transparent, and only the boundaries you want to remove are emitting, and the illumination surface is also "transparent", and not opaque, then you must take into account all the fixed T surfaces, as they will illuminate the "illuminator"* surface in a non-uniform way (try some visual ray tracing cases).

View factors can only be calculated if the geometries are defined and that COMSOL can estimate the relative viewed areas, and if I remember well (it's several months since I modelled with this radiative exchange), you need to define selectively the interacting surfaces that COMSOL should consider


--
Good luck
Ivar
Hi Probably I had a first sentence in my head, or it was a but ... from the middle of your long question ;) I see two/three cases: 1) your "illuminator" is infinite and opaque or covers 2*pi str above you illuminated area and your emissivity and diffusivities are fixed, known, or just =1 => you know the incident flux W/m^2 and you give an incident power density on the illuminated surface, and you forget about the illuminator domain 2) Or your illumination surface is limited to an area A_illu and you simplify the case and say the power emitted is fully absorbed, uniformly, (or partially if you can justify how much) over the illuminated Area and you say you have an incident flux density of the emitted power corrected for the relative Area ratio, and you forget about the illuminator domain too. 3) Or you have to take into account View Factors, hence solid angles, distances between surfaces, angles, Areas, emissivity, scatter ... And for that you must define the domain between the illuminator to the illuminated. still, if your illuminated domain is transparent, and only the boundaries you want to remove are emitting, and the illumination surface is also "transparent", and not opaque, then you must take into account all the fixed T surfaces, as they will illuminate the "illuminator"* surface in a non-uniform way (try some visual ray tracing cases). View factors can only be calculated if the geometries are defined and that COMSOL can estimate the relative viewed areas, and if I remember well (it's several months since I modelled with this radiative exchange), you need to define selectively the interacting surfaces that COMSOL should consider -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 27.08.2015, 15:38 GMT-4
Ivar,

I'm apparently not making a good job at explaining myself, and we have been talking about two different situations. But I think we are slowly getting to the point...

Forget your cases 1 and 2: if I could make such strong simplifications, I wouldn't need a FEA software like COMSOL! :-)

Let's concentrate on case 3:


3) Or you have to take into account View Factors, hence solid angles, distances between surfaces, angles, Areas, emissivity, scatter ...

Exaclty what I need! :-)


And for that you must define the domain between the illuminator to the illuminated.


No, that's not true. And it is where we are not understanding each other. If you don't care about the properties of the medium between the illuminator and the illuminated (that I don't... it's vacuum), then you do not need to model that domain, and all you view angles, distances, etc... are still correctly accounted for.
You don't even need to define the "containing domain" to account for the radiation towards ambient, as COMSOL (once you correctly set up the heat exchange interface) considers that in any direction that a surface is not seeing another surface, then it is seeing a "background" a Tamb.

Now again, the problem is: I have my domain representing the physical objects that are illuminated. I want to define a surface somewhere, at a fixed T>Tamb, that "illuminates" my object.
- If I create a domain (completely exterior to my illuminated objects, not "containing" them), then I can define one of its surfaces (looking towards the illuminated) as "emitting", and all works fine (see right panel on attached figure, in which I have colored the domain to make them easier to distinguish). However, I have the problem of having to include the additional domain, and its non-radiating faces, in the simulation.
- If I create just a surface, that would be ideal (left panel in the attached figure). Except, I do not know how to tell COMSOL to use that surface as a radiation source, since COMSOL only allows me to include domains, and not surfaces, in the "heat conduction" physics interface.

I hope it is now clear what my problem is...

Giacomo

Ivar, I'm apparently not making a good job at explaining myself, and we have been talking about two different situations. But I think we are slowly getting to the point... Forget your cases 1 and 2: if I could make such strong simplifications, I wouldn't need a FEA software like COMSOL! :-) Let's concentrate on case 3: [QUOTE] 3) Or you have to take into account View Factors, hence solid angles, distances between surfaces, angles, Areas, emissivity, scatter ... [/QUOTE] Exaclty what I need! :-) [QUOTE] And for that you must define the domain between the illuminator to the illuminated. [/QUOTE] No, that's not true. And it is where we are not understanding each other. If you don't care about the properties of the medium between the illuminator and the illuminated (that I don't... it's vacuum), then you do not need to model that domain, and all you view angles, distances, etc... are still correctly accounted for. You don't even need to define the "containing domain" to account for the radiation towards ambient, as COMSOL (once you correctly set up the heat exchange interface) considers that in any direction that a surface is not seeing another surface, then it is seeing a "background" a Tamb. Now again, the problem is: I have my domain representing the physical objects that are illuminated. I want to define a surface somewhere, at a fixed T>Tamb, that "illuminates" my object. - If I create a domain (completely exterior to my illuminated objects, not "containing" them), then I can define one of its surfaces (looking towards the illuminated) as "emitting", and all works fine (see right panel on attached figure, in which I have colored the domain to make them easier to distinguish). However, I have the problem of having to include the additional domain, and its non-radiating faces, in the simulation. - If I create just a surface, that would be ideal (left panel in the attached figure). Except, I do not know how to tell COMSOL to use that surface as a radiation source, since COMSOL only allows me to include domains, and not surfaces, in the "heat conduction" physics interface. I hope it is now clear what my problem is... Giacomo


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 28.08.2015, 03:08 GMT-4
Hi

Sorry for not understanding your issue at once, but there are so many possibilities ...

Probably what puzzles me is that what you are asking for is, for me, what the "Surface to Surface Radiation" module is exactly doing (or allows you to do)

I have tried to regenerate your problem in the model below in v5.1.0.180

check carefully the doc too, I see now that COMSOL has nicely improved the capabilities since I used it last :)

Here I'm not solving for an T, I impose them you need to add the HT for that (or use the full HT+Rad case

--
Good luck
Ivar
Hi Sorry for not understanding your issue at once, but there are so many possibilities ... Probably what puzzles me is that what you are asking for is, for me, what the "Surface to Surface Radiation" module is exactly doing (or allows you to do) I have tried to regenerate your problem in the model below in v5.1.0.180 check carefully the doc too, I see now that COMSOL has nicely improved the capabilities since I used it last :) Here I'm not solving for an T, I impose them you need to add the HT for that (or use the full HT+Rad case -- Good luck Ivar


Please login with a confirmed email address before reporting spam

Posted: 9 years ago 28.08.2015, 15:10 GMT-4
Hi Ivar,

Unfortunately I cannot open the model you attached, as I'm limited to COMSOL 5.0. However, by looking at the screenshot, I do have a comment: I don't think you need to insert the two extra domains (triangle and trapezoid) besides the two rectangles.

Anyway, I think we are finally hitting the right spot. I am attaching two simple model files similar to yours (I should have done this from the beginning):

- The first one, "RadiationExample - Solid.mph", has two rectangular domains, one for the illuminated and one for the illuminator. It uses a single physics interface, ht, with the "surface-to-surface" radiation option enabled. This model does almost exactly what I want, except for one thing, that is the object of my original question: I would like to get rid of the domain representing the illuminator, as I really only care about the emitting boundary.

- "RadiationExample - Surface.mph" is an attempt at accomplishing what I stated above. I first deleted the domain representing the illuminator and replaced it with just the emitting boundary. However, I cannot include the boundary itself in the HT physics interface, since it is not part of a domain.
So, rather than enabling the option in the HT module, I added a separate "Surface-to-Surface" physics interface. This allows me to include only boundaries. However, I have a problem coupling the two: I can tell the radiation interface to use the temperature computed by the ht interface to calculate the radiation, but I haven't found the way of telling the ht interface to use the radiation as a boundary heat flux.

Where to go from here? :-)

Giacomo

Hi Ivar, Unfortunately I cannot open the model you attached, as I'm limited to COMSOL 5.0. However, by looking at the screenshot, I do have a comment: I don't think you need to insert the two extra domains (triangle and trapezoid) besides the two rectangles. Anyway, I think we are finally hitting the right spot. I am attaching two simple model files similar to yours (I should have done this from the beginning): - The first one, "RadiationExample - Solid.mph", has two rectangular domains, one for the illuminated and one for the illuminator. It uses a single physics interface, ht, with the "surface-to-surface" radiation option enabled. This model does almost exactly what I want, except for one thing, that is the object of my original question: I would like to get rid of the domain representing the illuminator, as I really only care about the emitting boundary. - "RadiationExample - Surface.mph" is an attempt at accomplishing what I stated above. I first deleted the domain representing the illuminator and replaced it with just the emitting boundary. However, I cannot include the boundary itself in the HT physics interface, since it is not part of a domain. So, rather than enabling the option in the HT module, I added a separate "Surface-to-Surface" physics interface. This allows me to include only boundaries. However, I have a problem coupling the two: I can tell the radiation interface to use the temperature computed by the ht interface to calculate the radiation, but I haven't found the way of telling the ht interface to use the radiation as a boundary heat flux. Where to go from here? :-) Giacomo

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.